Authorship joins recognition to accountability
A byline does more than acknowledge participation. It allocates scholarly credit and identifies people who can answer for named parts of the work. Hierarchy or membership of a laboratory is therefore insufficient. Conversely, a contributor whose work qualifies should not be excluded for tactical or institutional reasons.
Disciplines and journals differ in order and detail. Read the target journal's policy, field conventions, institutional rules and relevant ICMJE guidance before deciding. Apply the same stated test to everyone. This page supports good research practice; it is not contractual or legal advice.
Apply four connected authorship conditions
The widely used ICMJE framework requires a substantial contribution to conception, data or analysis; participation in drafting or critical intellectual revision; approval of the version to be published; and willingness to be accountable for the work and help resolve integrity questions. The four conditions operate together.
A person who meets the first condition should have a fair opportunity to meet the remaining ones. Deliberately excluding them from drafting or approval must not be used to prevent qualification. Approval needs adequate time and access to the relevant manuscript, not a last-minute request for a ceremonial yes.
Accountability has identifiable parts without becoming isolated. Each author should know their own work in depth and understand how confidence in other parts was established. An unexplained dataset or inaccessible analysis cannot simply be assigned to someone who is absent.
Use CRediT to describe work, not award status
CRediT names fourteen roles including conceptualization, methodology, software, validation, analysis, investigation, resources, data curation, writing, visualisation, supervision, administration and funding acquisition. It reveals the breadth of research labour. Several contributors may share a role at different levels.
The taxonomy deliberately does not prescribe an authorship threshold or byline order. Funding or administration may be vital but does not alone satisfy every authorship condition. Work below the threshold can be recognised, with consent, in acknowledgements or a contributor statement. Specialist services should be described accurately.
Write contributions with verbs and objects. “Reviewed the coding of 40 interviews” is more informative than “Validation”. Avoid personality judgements. Update the record when actual responsibilities change rather than preserving an obsolete plan.
Revisit the decision at defined milestones
- At project start, agree the criteria, expected roles and review points.
- After major data collection or analysis, update the matrix from actual work.
- Before drafting, discuss the author group and provisional order.
- Before submission, confirm contributions, interests, approval and accountability.
- After substantial revision, record new work and reconsider changes.
Minutes should include date, participants, decision, reasons and next review. They support rather than replace a conversation. A surprise order change near submission deserves an explicit explanation grounded in actual contribution.
Resolve disagreements through evidence, not power
Ask each contributor to set out specific work and supporting records. Compare it with the rule agreed earlier and the journal policy. A neutral ombuds or research-integrity office can facilitate an unresolved dispute. No one should be coerced through dependency, threats or withheld data.
The CRediT guide examines roles in detail. The three-case comparison identifies improper attribution. More resources sit in the publication ethics hub. The single foundation route is the guide to reviewing documentary evidence.
Address exceptional cases explicitly. A deceased or unavailable contributor may still qualify under the relevant policy where contribution and accountability arrangements are documented; an institution does not simply inherit personal authorship. Group authorship needs a published membership and responsibility structure. Large collaborations may use a byline, consortium name and contributor list for different purposes, but readers should still be able to identify work and a route for questions.
Version the matrix with the manuscript. This shows which roles applied at first submission and what changed during a substantial revision. New analysis, added data collection or major redrafting may alter attribution, whereas minor formatting does not automatically do so. Record the reason rather than fitting history to a preferred order.